Assignment 1 - ESM 244
Claire Madden
2/6/2020
Task 1.
Prepare a single polished HTML (knitted from .Rmd), planning that this might be a post/project you’d include on your personal blogdown site, that includes at least:
A useful descriptive introductory summary (3 - 4 sentences) of what’s contained in the project
Image source: National Geographic
All of your organized and well-annotated code (with warnings/messages hidden) used to create at least:
One finalized graph about the Bonanza Creek snowshoe hare population (you pick which variables, how you want to wrangle it beforehand, and which type of visual to create - but make sure it is beautifully finalized)
One finalized HTML table (probably created using kable & kableExtra) containing summary statistics about the snowshoe hares (again, you pick which variables, and how you want to group/summarize them)
Make sure that both your figure and table appear in your final knitted document, each with a useful caption. Include text associated with each to help the audience understand and interpret the results.
library(tidyverse)
library(kableExtra)
library(janitor)
library(lubridate)
library(tidyr)
library(ggfortify)
library(RColorBrewer)
hares<- read_csv("showshoe_lter.csv")
hares_mf <- hares %>%
mutate(sex = str_to_lower(sex)) %>% # change all entries in sex column to lower case
filter(sex == "m" | sex == "f") %>% # remove all observations with NA or ? in sex
mutate(date = mdy(date)) %>% # change date format
mutate(full_date = date) %>% # retain date column
separate(date, into = c("year", "month", "day")) %>% #create new columns with parsed date info
mutate(month = as.numeric(month)) %>%
mutate(season = ifelse(month %in% c(1:4, 11, 12), "winter", "summer")) %>% # organize months into "summer" and "winter"
mutate(year = as.numeric(year))
# seems like there are a lot more observations in the summer than in the winter so maybe not an interesting thing to use for visualization
# explore graph
ggplot(data = hares_mf, aes(x = weight, y = hindft))+
geom_point()+
facet_wrap(~sex)
# this is not that interesting to me# try something else that may be more interesting
hares_hist <- ggplot(data = hares_mf, aes(x = year))+
geom_bar(aes(fill = sex), position = "dodge")+
scale_fill_manual(values = c("peru", "peachpuff2"), labels = c("Female", "Male"))+
scale_x_continuous(expand = c(0,0),
breaks = seq(1998, 2012, 1))+
theme_minimal()+
labs(x = "Year of Observation",
y = "Observation Count")+
theme(legend.title = element_blank())
hares_hist
Figure 1. Snowshoe Hare Monitoring Observations in the Bonanza Creek Experimental Forest Between 1998 - 2012.
Total number of observations are recorded each year during the monitoring period and differentiated by sex.
Data source: Kielland K., F. S. Chapin, R. W. Ruess. 2017.
# make a summary table for females
hare_summary_f <- hares_mf %>%
filter(sex == "f") %>%
drop_na(weight, hindft) %>% # take out an values to be able to calculate mean
group_by(year) %>% # organize by year and sex
summarize(
count = length(sex), # number of observations
avg_weight = round(mean(weight), digits = 0))# mean weight
# make a summary table for males
hare_summary_m <- hares_mf %>%
filter(sex == "m") %>%
drop_na(weight, hindft) %>% # take out an values to be able to calculate mean
group_by(year) %>% # organize by year and sex
summarize(
count = length(sex), # number of observations
avg_weight = round(mean(weight), digits = 0)) # mean weight
# combine the two tables!
hare_summary <- merge(hare_summary_f, hare_summary_m, by = "year")
# format into a single pretty table
hare_table <- kable(hare_summary, align = "c", col.names = c("Year", "Count", "Average Weight", "Count", "Average Weight"), caption = "Table 1. Sumamry Statistics for Snowshoe Hare Monitoring Observations at Bonanza Creek Experimental Forest Between 1998 - 2012.") %>%
add_header_above(c(" " = 1, "Female" = 2, "Male" = 2)) %>%
kable_styling(bootstrap_options = c("striped"), full_width = FALSE)
hare_table| Year | Count | Average Weight | Count | Average Weight |
|---|---|---|---|---|
| 1998 | 10 | 1746 | 5 | 1678 |
| 1999 | 113 | 1335 | 136 | 1241 |
| 2000 | 78 | 1517 | 47 | 1302 |
| 2001 | 25 | 1482 | 27 | 1420 |
| 2002 | 10 | 1014 | 17 | 1301 |
| 2003 | 33 | 1206 | 23 | 1223 |
| 2004 | 45 | 1536 | 29 | 1372 |
| 2005 | 47 | 1162 | 36 | 1297 |
| 2006 | 29 | 1283 | 15 | 1443 |
| 2007 | 57 | 1412 | 25 | 1190 |
| 2008 | 87 | 1331 | 76 | 1360 |
| 2009 | 117 | 1281 | 100 | 1325 |
| 2010 | 96 | 1299 | 43 | 1350 |
| 2011 | 85 | 1339 | 43 | 1400 |
| 2012 | 24 | 1291 | 21 | 1331 |
Data source: Kielland K., F. S. Chapin, R. W. Ruess. 2017. Snowshoe hare physical data in Bonanza Creek Experimental Forest: 1999-Present. Environmental Data Initiative. https://doi.org/10.6073/pasta/03dce4856d79b91557d8e6ce2cbcdc14.
Task 2.
This dataset, compiled and provided by @zander_venter on Kaggle, is a compilation of data aquired through Google Earth Engine (https://earthengine.google.com/). The dataset uses publically available remote sensing data, where most of the data is derived by calculating the mean for each country at a scale of about 10km. In order to understand the relationship between different environmental and climatic variables included in the dataset, a principal components analysis is provided below to demonstrate directionality of the greatest source of variance within variables considered. These variables were chosen to demonstrate the relationship of variance between different climatic variables included in the dataset.
Takeaways from the principal components analysis: - Elevation and temperature are negatively correlated - Isothermality and temperature are positively correlated - Mean annual rain, tree cover, and cloudiness are all positively correlated - Mean annual rain, tree cover, and cloudiness are not related to elevation or temperature - Principal components 1 and 2 explain 75% of variance in the data
# explore data
world_env <- read_csv("world_env_vars.csv") %>%
clean_names(case = "upper_camel")
summary(world_env) # check out the dataset# select variables of interest for principal components analysis
env_subset <- world_env %>%
select(Elevation, TreeCanopyCover, Isothermality, RainMeanAnnual, TempMeanAnnual, Cloudiness)
# explore missingness
VIM::matrixplot(env_subset, sortby = "Elevation")
# can't see variable names so not that helpful for exploration
naniar::gg_miss_var(env_subset)
# only a few NAs for each variable, will remove them
env_nona <- env_subset %>%
drop_na() # remove NA values
env_pca <- prcomp(env_nona, scale = TRUE) # run PCA
plot(env_pca)
biplot(env_pca) # lookin real UGLY! 
# a prettier version...
my_pca_plot <- autoplot(env_pca,
colour = NA,
loadings.colour = "black",
loadings.label = TRUE,
loadings.label.size = 3,
loadings.label.colour = "black",
loadings.label.repel = TRUE)+
theme_bw()
my_pca_plot
Figure 2. Principal components analysis showing the relationship between variables included in a compilation
of Google Earth Engine datasets.
Data source: @zander_venter.